{chem}[foss/2020b] Dalton v2020.0 #13921

LHurst-JM · 2021-09-07T10:08:32Z

Bump of Dalton from 2016 to 2020.0 and build for foss-2020b (only other version in easybuild is for intel-2017b).

Dalton's new (python) configure command seems to fix the issues that needed a lot of patches in 2016.

N.B. This eb takes between 3 and 4.5 hours to build on my cluster (on each of 44-core Skylake, 60-core Zen and 120-core Zen2 nodes). Most of that time is spent running the test suite. One test fails on Zen2 due to upstream bug - https://gitlab.com/dalton/dalton/-/issues/183

…make path is also incorrect

…onfigs into dalton

…asyconfigs into dalton

boegelbot · 2021-09-07T10:53:29Z

@LHurst-JM: Tests failed in GitHub Actions, see https://github.com/easybuilders/easybuild-easyconfigs/actions/runs/1208994440
Output from first failing test suite run:

======================================================================
FAIL: test_pr_sha256_checksums (test.easyconfigs.easyconfigs.EasyConfigTest)
Make sure changed easyconfigs have SHA256 checksums in place.
----------------------------------------------------------------------
Traceback (most recent call last):
  File "test/easyconfigs/easyconfigs.py", line 843, in test_pr_sha256_checksums
    self.assertTrue(len(checksum_issues) == 0, "No checksum issues:\n%s" % '\n'.join(checksum_issues))
AssertionError: No checksum issues:
Checksums missing for one or more sources/patches in Dalton-2020.0-foss-2020b.eb: found 1 sources + 0 patches vs 0 checksums

----------------------------------------------------------------------
Ran 13281 tests in 484.026s

FAILED (failures=1)
Merge base for develop and HEAD: dd7d938a15cd5e4cd9b28e1594c0b24084f52455
Merge base for develop and HEAD: dd7d938a15cd5e4cd9b28e1594c0b24084f52455

List of added easyconfig files in this PR:
	Dalton-2020.0-foss-2020b.eb
ERROR: Not all tests were successful

bleep, bloop, I'm just a bot (boegelbot v20200716.01)
Please talk to my owner @boegel if you notice you me acting stupid),
or submit a pull request to https://github.com/boegel/boegelbot fix the problem.

verdurin · 2021-09-13T09:04:54Z

@boegelbot please test @ generoso

boegelbot · 2021-09-13T09:05:09Z

@verdurin: Request for testing this PR well received on login1

PR test command 'EB_PR=13921 EB_ARGS= /opt/software/slurm/bin/sbatch --job-name test_PR_13921 --ntasks=4 ~/boegelbot/eb_from_pr_upload_generoso.sh' executed!

exit code: 0
output:

Submitted batch job 6878

Test results coming soon (I hope)...

- notification for comment with ID 917988119 processed

Message to humans: this is just bookkeeping information for me,
it is of no use to you (unless you think I have a bug, which I don't).

boegelbot · 2021-09-13T09:06:54Z

Test report by @boegelbot
FAILED
Build succeeded for 0 out of 1 (1 easyconfigs in total)
cnx1 - Linux rocky linux 8.4, x86_64, Intel(R) Xeon(R) CPU E5-2690 v3 @ 2.60GHz (haswell), Python 3.6.8
See https://gist.github.com/042009d5dca6026b0c148055fcb4bb95 for a full test report.

LHurst-JM · 2021-09-13T09:20:28Z

Checksum failed on the source. That's definitely the checksum for the source that was used in my testing:

$ sha256sum /mnt/devtest-apps/easybuild/2020b/sources/d/Dalton/Dalton-2020.0.tar.gz 
a939dfbf869ef90ea064fa3d00cc1937dd145438947ecf5556f964b38299b34a  /mnt/devtest-apps/easybuild/2020b/sources/d/Dalton/Dalton-2020.0.tar.gz

Upstream have not changed the commit the tag points to:

 Dalton2020.0 protected
 66052b3a · Merge branch 'bump-version' into 'release/2020' · 10 months ago
 Release Dalton2020.0

So I'm not sure why the checksum is different. I'll delete and re-download the source with --fetch...

LHurst-JM · 2021-09-13T10:16:50Z

Somehow the source checksum has changed upstream since I uploaded the easyconfig. The tag still points to the same commit, so I really don't understand why?:

$ sha256sum /mnt/devtest-apps/easybuild/2020b/sources/d/Dalton/Dalton-2020.0.tar.gz 
5f37133fe9e791351a6578edc62b56f11cd9512344e9d30401cd9d2cac91c848  /mnt/devtest-apps/easybuild/2020b/sources/d/Dalton/Dalton-2020.0.tar.gz

verdurin · 2021-09-13T16:26:28Z

Is there any reason not to use one of the upstream tarballs instead, e.g. https://gitlab.com/dalton/dalton/-/archive/Dalton2020.0/dalton-Dalton2020.0.tar.bz2 ?

LHurst-JM · 2021-09-13T16:45:13Z

Is there any reason not to use one of the upstream tarballs instead, e.g. https://gitlab.com/dalton/dalton/-/archive/Dalton2020.0/dalton-Dalton2020.0.tar.bz2 ?

Only that I did not know they existed ;)

Upstream's website's download link takes you to the 'release/2020' branch HEAD, which is under active development (and therefore unsutable to use as the source). The tag seems stable and this is exactly how Dalton-2016-intel-2017b-i8.eb, which is the only version currently in EB, fetches it. I'm still very confused by the tag, which allegedly hasn't changed in 10 months, suddenly resulting in a different checksum for the clones source.

Happy to try the one of those archive links instead.

LHurst-JM · 2021-09-13T16:48:12Z

...so that results in a file with yet another checksum:

$ wget https://gitlab.com/dalton/dalton/-/archive/Dalton2020.0/dalton-Dalton2020.0.tar.gz
$ sha256sum dalton-Dalton2020.0.tar.gz 
6b8365874af6c1545db98f6deb0decce97f6a689995b1cc2a57fea4efd399f00  dalton-Dalton2020.0.tar.gz

(via the releases page, which claims this source file is the same commit as the tag)

🤦

akesandgren · 2021-10-14T15:30:10Z

See #14100 and #14154 for missing patches

akesandgren · 2021-10-14T15:38:59Z

Also you need recursive download here to get the full functionality and can thus not use the simple tarball.

akesandgren · 2021-10-14T15:43:43Z

easybuild/easyconfigs/d/Dalton/Dalton-2020.0-foss-2020b.eb

+# 'runtests'
+buildopts = "&& ctest --output-on-failure -L dalton"
+
+separate_build_dir = True


This is default for CMakeMake and can be removed

It doesn't work without this (-L dalton) in my environment?

This is based on how other CMakeMake easyconfigs work.

LHurst-JM · 2021-10-15T09:01:04Z

Also you need recursive download here to get the full functionality and can thus not use the simple tarball.

You don't if you seperate the extrenal dependencies into their own ebs.

akesandgren · 2021-10-15T12:28:17Z

Yes, that would be preferable, splitting the submodules out that is, but I haven't had time to look at that...

Also look at #14154 and #14158
That is how I did things to make it work with tests.
Note that all the patches are necessary, but for pre-2021a toolchains, the flexiblas patch is not needed.

I'm currently running that one with tests on broadwell and zen2

akesandgren · 2021-10-15T13:23:41Z

I get the same error on my zen2 system that you saw with my foss-2021a build.

akesandgren · 2021-10-15T17:19:41Z

And see the #14158 for the final details on how to run the tests and the required patches.

boegel · 2024-01-13T18:57:32Z

closing this since foss/2020b is deprecated, see https://docs.easybuild.io/policies/toolchains

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

{chem}[foss/2020b] Dalton v2020.0 #13921

{chem}[foss/2020b] Dalton v2020.0 #13921

LHurst-JM commented Sep 7, 2021 •

edited

Loading

boegelbot commented Sep 7, 2021

verdurin commented Sep 13, 2021

boegelbot commented Sep 13, 2021

boegelbot commented Sep 13, 2021

LHurst-JM commented Sep 13, 2021

LHurst-JM commented Sep 13, 2021

verdurin commented Sep 13, 2021

LHurst-JM commented Sep 13, 2021

LHurst-JM commented Sep 13, 2021 •

edited

Loading

akesandgren commented Oct 14, 2021

akesandgren commented Oct 14, 2021

akesandgren Oct 14, 2021

LHurst-JM Oct 15, 2021

LHurst-JM commented Oct 15, 2021

akesandgren commented Oct 15, 2021

akesandgren commented Oct 15, 2021

akesandgren commented Oct 15, 2021

boegel commented Jan 13, 2024

{chem}[foss/2020b] Dalton v2020.0 #13921

{chem}[foss/2020b] Dalton v2020.0 #13921

Conversation

LHurst-JM commented Sep 7, 2021 • edited Loading

boegelbot commented Sep 7, 2021

verdurin commented Sep 13, 2021

boegelbot commented Sep 13, 2021

boegelbot commented Sep 13, 2021

LHurst-JM commented Sep 13, 2021

LHurst-JM commented Sep 13, 2021

verdurin commented Sep 13, 2021

LHurst-JM commented Sep 13, 2021

LHurst-JM commented Sep 13, 2021 • edited Loading

akesandgren commented Oct 14, 2021

akesandgren commented Oct 14, 2021

akesandgren Oct 14, 2021

Choose a reason for hiding this comment

LHurst-JM Oct 15, 2021

Choose a reason for hiding this comment

LHurst-JM commented Oct 15, 2021

akesandgren commented Oct 15, 2021

akesandgren commented Oct 15, 2021

akesandgren commented Oct 15, 2021

boegel commented Jan 13, 2024

LHurst-JM commented Sep 7, 2021 •

edited

Loading

LHurst-JM commented Sep 13, 2021 •

edited

Loading